get_annotations_attributes

DetectionEvaluator.get_annotations_attributes(predictions_name: str | None = None) → list[str]

Get the name of columns related to annotations attributes. In other words, get columns that are NOT the default ones.

the actual attribute values can then be

self.predictions_dictionary[predictions_name][
    self.get_annotations_attributes()
]

Parameters:

predictions_name – name of predictions to extract not default column from. If None, will use self.groundtruth. Defaults to None.

Returns:

list of column names in self.annotations that represent attributes